Building Knowledge Graphs With LLMs
summary
Shows a practical workflow for turning PDFs into Neo4j knowledge graphs using LLMs with concrete trade offs: use generic→schema→domain-specific extraction depending on effort and accuracy, prefer chunk size matters and multi‑pass extraction to increase recall, and apply metadata filtering when aggregations like total contract value must be exact. Demonstrates tooling by name—LangChain, LlamaIndex, OpenAI structured output, Knowledge Graph Builder and Neo4j AuraDB—and stresses postprocessing steps like entity resolution to merge duplicates and make queries and aggregations reliable.
AI-written summary · about
chapters (10)
- 0:00Limitations of RAG
- 1:02When naive RAG fails
- 6:18Introduction to KG
- 7:51KG and data representation
- 10:28Information extraction basics
- 14:32Generic extraction approach
- 17:21Middle-ground extraction
- 19:39Domain-specific extraction
- 23:07Observations in practice
- 27:18Questions and answers
related talks
20mPractical GraphRAG: Making LLMs smarter with Knowledge Graphs
44mGive Your LLMs a Left Brain
53mPractical GraphRAG - Making LLMs smarter with Knowledge Graphs
28mGrounded Intelligence Boosting LLMs with Real-Time Graph Context
23mKnowledge Graphs and LLMs in Action
26mPractical GraphRAG Making LLMs smarter with Knowledge Graphs
41mEnhancing LLMs with Graph Technology
52mGive Your LLMs a Left Brain